Abstract

Should We Trust a Self-Driving Vehicles?

Author(s): Luc Rolland

Statement of the Problem: The fantastic achievements in the soft- and hardware sector in the last few years are a key enabler in conservative industries to apply new technologies. In the automotive industry, for example, the race for a self-driving vehicle has already started to come up with a driverless system. Artificial intelligence (AI) and new sensor technologies increase the development speed and
open new ways to handle very complex traffic scenarios on the level of a human. In controlled environments like freeways, a few systems have already reached an excellent safety and accuracy level. Adaptive Cruise Control (ACC) and Lane Keeping Assistant (LKA) systems, for instance, are already trusted and accepted by drivers. However, there is also a big problem coming with this trust because the assistance systems allow the driver to perform different activities like writing emails during a drive. Today’s system features are not capable of handling all edge problems. These limitations can lead to fatal traffic accidents because drivers are not aware of the feature maturity and the consequences.